[M3devel] win32 threads...now Juno sometimes hangs..

[Jay K]

ah, well now I've got 68 runs with an event amd /slight/ expansion of the giant lock -- in InnerWait I .release before Leave(giant).

Given that release almost immediates enter giant anyway, that doesn't seem a bad move.

 

It didn't hang at 68, it did:

 

***
*** runtime error:
***    An array subscript was out of range.
***    file "..\src\vbt\VBTRep.m3", line 644
***

Stack trace:
   FP         PC      Procedure
---------  ---------  -------------------------------
0x207fef8   0xfb62a5  Redisplay + 0x3ae in ..\src\vbt\VBTRep.m3
0x207ff24   0xfb5e34  UncoverRedisplay + 0xdd in ..\src\vbt\VBTRep.m3
0x207ff4c   0xfb5ec5  RdApply + 0x8c in ..\src\vbt\VBTRep.m3
0x207ff8c   0x5eacb9  RunThread + 0x184 in ..\src\thread\WIN32\ThreadWin32.m3
0x207ffb4   0x5eaaea  ThreadBase + 0x33 in ..\src\thread\WIN32\ThreadWin32.m3
.........  .........  ... more frames ...

C:\dev2\cm3.2\m3-libs\m3core>echo 69   && \cm3\bin\Juno.exe @M3no-trestle-await-
delete
69

C:\dev2\cm3.2\m3-libs\m3core>echo 70   && \cm3\bin\Juno.exe @M3no-trestle-await-
delete
70

 

I'll try again having put back the semaphore..

I rather suspect I have it fixed though and have uncovered some other problem.

 

 - Jay
 


From: jay.krell at cornell.edu
To: hosking at cs.purdue.edu
Date: Tue, 29 Sep 2009 14:48:24 +0000
CC: m3devel at elegosoft.com
Subject: Re: [M3devel] win32 threads...now Juno sometimes hangs..



Tony..this semaphore with a maximum value of 1, that is a strange thing, isn't it?
It could just be an event, right?
I tried that..it took 50 runs to hang Juno..
 
 - Jay
 


From: jay.krell at cornell.edu
To: hosking at cs.purdue.edu
CC: m3devel at elegosoft.com
Subject: RE: [M3devel] win32 threads...now Juno sometimes hangs..
Date: Tue, 29 Sep 2009 14:23:07 +0000



ps: while I consider the whole stacksize thing broken anyway, notice that using an idle thread (which you had no choice about) got you an indeterminate stack size.
 
 - Jay



From: jay.krell at cornell.edu
To: hosking at cs.purdue.edu
Date: Tue, 29 Sep 2009 14:16:11 +0000
CC: m3devel at elegosoft.com
Subject: Re: [M3devel] win32 threads...now Juno sometimes hangs..



small email truncation:
 
>> We need some sort of stress or fault-injection or variation-injection tests.
>> Run threads deterministically in every combination of order, for example.


 - Jay

 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20090929/0a7c5a31/attachment-0002.html>